Log file Viewer

[16:10:03] <Ange7> I will bench it

[16:10:06] <StephenLynx> what version you tested, btw?

[16:10:07] <Derick> StephenLynx: because that's what MongoDB wants

[16:11:04] <StephenLynx> I can serve full html pages straight from mongo in 2ms.

[16:11:12] <Derick> i'm sure you can

[16:11:26] <StephenLynx> your benchmark was probably botched.

[16:11:28] <Derick> but we are benchmarking raw driver performance

[16:11:36] <Derick> StephenLynx: the node.js driver engineer wrote it

[16:11:43] <StephenLynx> v:

[16:12:07] <StephenLynx> id like to see the whole of these benchmarkes.

[16:12:14] <cheeser> Derick: Val?

[16:12:21] <Derick> cheeser: I thought Christian

[16:12:26] <cheeser> oh, right!

[16:12:42] <cheeser> i was trying to remember who was on those bits.

[16:12:45] <cheeser> poor sods.

[16:13:08] <Derick> cheeser: see how I haven't mentioned how my driver performed?

[16:13:25] <cheeser> Derick: save that crow for another day. ;)

[16:21:30] <geri> hi how can i query all type1? http://ideone.com/uyju5J

[16:22:02] <cheeser> { "SIO.type1": ...}

[16:24:05] <geri> what is ... ?

[16:24:18] <geri> i want to read all SIO.type1

[16:25:11] <cheeser> whatever your query criteria is.

[16:25:33] <geri> cheeser: i want to select all existing

[16:25:52] <cheeser> they'll be part of the docs you get back from your query...

[16:27:16] <geri> cheeser: im not sure what you mean?

[16:28:11] <geri> cheeser: dont you need to set the criteria before you run the query?

[16:28:54] <cheeser> you can query with {}

[16:30:17] <geri> cheeser: you mean: { "SIO.type1": "{}"} ...?

[16:30:33] <cheeser> no, i meant {}

[16:30:42] <cheeser> that will match every document in your collection

[16:30:54] <geri> cheeser: i only want to get the SIO.type1 data

[16:31:09] <cheeser> https://docs.mongodb.org/manual/tutorial/project-fields-from-query-results/

[16:53:57] <geri> cheeser: i used: { "SIO.type1": { "$exists": true}} ... but it does not return the only return the type1 fields...any idea?

[17:24:09] <geri> cheeser: i used: { "SIO.type1": { "$exists": true}} ... but it does not return the only return the type1 fields...any idea?

[17:39:29] <cheeser> geri: https://docs.mongodb.org/manual/reference/operator/query/exists/

[17:41:32] <geri> cheeser: yes?

[17:42:27] <geri> cheeser: is that simply the wrong operation?

[17:42:37] <cheeser> it's the wrong order

[17:43:13] <geri> cheeser: can you explain?

[17:43:47] <cheeser> look at your query. look at the docs.

[17:45:51] <geri> "$exists matches the documents that contain the field" ... the field is "SIO.type1"

[17:51:16] <geri> cheeser: "$exists matches the documents that contain the field" ... the field is "SIO.type1"

[17:52:15] <cheeser> yes. look at the order of elements in your query. compare with the order in the docs.

[17:53:24] <geri> cheeser: i dont see an issue with the order...

[17:55:49] <geri> cheeser: the docs writes: Syntax: { field: { $exists: <boolean> } }

[17:55:55] <geri> cheeser: what is the diff?

[17:56:05] <cheeser> oh, whoops! haha. i totally misread that.

[17:56:17] <cheeser> my bad. it had scrolled off screen and was misremembering it.

[17:56:18] <phishy> New to map/reduce and I'm pretty sure I can't get what I want using it here, but wondering. Given this data structure, is it possible to retrieve a set of usage over time for a single user? https://gist.github.com/phishy/2be374cfb9acea05d1823615cf3eed46

[17:56:41] <geri> cheeser: i expect the result to return only SIO.type1 data...

[17:57:05] <geri> cheeser: but instead it returns the everything... any idea?

[17:57:18] <geri> cheeser: but instead it returns everything... any idea?

[17:57:20] <cheeser> that's how queries work. if you want a subset, you have to use a projection.

[17:57:39] <cheeser> 12:24 <@cheeser> https://docs.mongodb.org/manual/tutorial/project-fields-from-query-results/

[17:57:40] <geri> cheeser: this is done after the query?

[18:10:31] <geri> cheeser: ok i see, now i still get the msg projection cannot have a mix of inclusion and exclusion...

[18:11:35] <geri> so it needs to set the exclusion before?

[18:13:17] <cheeser> set the exclusion before?

[18:13:25] <cheeser> i don't know what that means.

[18:14:37] <geri> cheeser: i mean define the exclusion before the inclusion

[18:18:09] <cheeser> you can't mix them. either include or exclude.

[18:21:15] <StephenLynx> UNLESS

[18:21:20] <StephenLynx> you are excluding _id

[18:21:45] <StephenLynx> because of course, _id is a very special snowflake and by no means just another field.

[19:44:40] <phishy> Is it possible to query in mongo if a record has a certain object key?

[19:44:57] <cheeser> yes. $exists

[19:47:03] <phishy> cheeser: I thought that checked if the document had a certain field?

[19:55:51] <cheeser> you give it a key name and $exists will tell you if that key is there or not.

[20:32:36] <jr3> is there any way of $inc a document field by an amount specified on the doc?

[20:32:58] <jr3> instead of coming from app code?

[20:33:11] <Derick> not that I know of

[20:53:13] <phil22> hello

[20:53:27] <phil22> I'm trying to understand why a query is extremely slow

[20:53:36] <phil22> here is the profiling of the query:

[20:53:38] <phil22> http://pastebin.com/BUwW9r21

[20:54:27] <cheeser> and the query?

[20:57:01] <phil22> cheeser: the query is .find(phoneValid:true).orderBy(date,-1).limit(15)

[22:07:21] <Waheedi> finally switch to 3.2 :S

[22:07:33] <cheeser> w00t

[22:12:34] <Waheedi> WiredTiger :)

[22:12:39] <Waheedi> what a name

[22:13:28] <Waheedi> THE BROWN BEAR IS COMING YOUR WAY :P

[22:36:17] <GothAlice> Huh. https://docs.mongodb.org/ecosystem/platforms/rackspace-cloud/#topology < why is the documentation suggesting a setup with an even number of voting members? (Third graphic, "hybrid".)

[22:40:33] <cheeser> hrm. how (not) odd.

[22:41:28] <cheeser> probably just a space/size issue on the graphic

[22:52:10] <GothAlice> cheeser: A confused user I'm helping elsewhere asked the astute question "uh, should… should I add an arbiter to this?" which a) made me very happy that he's learning, and b) worried that the documentation ought not suggest things (even casually) that will definitely lead to problems. ;)

[22:54:41] <uuanton> anyone know how to set blockdev --setra 32 /dev/xvdf on start on centos 7 ?

[22:55:00] <uuanton> somewhere on /lib/systemd/system/mongod.service

[22:58:27] <uuanton> i have warning about WARNING: Readahead for /data is set to 4096KB

[23:04:42] <uuanton> i hate the day when we upgraded to centos7

[23:06:01] <uuanton> settings inside "/etc/rc.local" and "/etc/udev/rules.d/85-ebs.rules" doesnt help

[23:06:10] <MacWinner> does wiretiger give you the benefit of document level locking vs mmap in 3.0.x?

[23:42:33] <MacWinner> do indexes with wiredtiger need to be able to fit in memory?

[23:44:27] <Boomtime> anything you access frequently will be better off in memory, this is true of wiredtiger and mmap - also, neither of those have any requirement for an index to entirely fit in memory, doing so is a performance concern only